The Natural Sciences and Engineering Research Council of Canada has granted $498,885 to Queen's University law professor Samuel Dahan for a five-year international study aimed at reshaping how courts use artificial intelligence. The project arrives as 78% of Canadians facing significant legal issues cannot secure legal assistance, pushing more self-represented litigants into court systems already strained by AI-generated filing errors.
Dahan, founder of the open-source platform OpenJustice, will serve as principal investigator of "Courtroom Intelligence: Co-Designing AI for Courts." The initiative convenes experts in law, computer science, linguistics, and social science alongside judicial partners from multiple countries. Queen's Smith School of Business and the Department of Electrical and Computer Engineering will participate, along with Toronto Metropolitan University, McGill University, Harvard Law School, University College London, and Leiden University in partnership with the Council of Europe. Former Supreme Court of Canada Justice Thomas Cromwell, the French Cour de cassation, and the Hong Kong Judiciary will also contribute.
A co-designed approach with judicial partners
"This project is unique because it's truly co-designed with courts. We are not building technology in isolation; we are working directly with judges, clerks, and legal practitioners to ensure it meets real-world needs," Dahan said. The funds, distributed through NSERC's Discovery Horizons program, will support the development of an open-source judicial AI framework tailored specifically to court environments.
Many self-represented litigants have turned to AI tools for help with legal filings, but existing systems often hallucinate and produce error-filled documents that delay proceedings and waste judicial resources. Dahan's approach targets a gap in AI for Legal applications: most existing tools are not designed for the realities of judicial reasoning.
Building inspectable AI reasoning
"AI has enormous potential to support access to justice. But today's systems are not designed for the realities of legal reasoning or judicial work. Our goal is to build AI systems whose reasoning courts can inspect, evaluate, and trust - systems grounded in how legal decisions are made, not just in large volumes of text," Dahan said.
Central to the project is a reasoning flows no-code framework that lets legal professionals construct detailed representations of legal analysis through a visual interface. Judges and lawyers can model processes like statute identification, evidence evaluation, and precedent application directly within the system. Pilot studies have shown this approach yields fewer errors than traditional AI systems. "[Judges and lawyers] can encode their reasoning directly into the system, ensuring that outputs are explainable, auditable, and aligned with legal standards," Dahan said.
The project will also concentrate on drafting assistance, decision standardization, evidence admissibility, and deepfake detection. Researchers will examine methods to help courts identify synthetic and manipulated evidence, a growing concern as generative AI becomes more sophisticated.
Training the next generation and expanding infrastructure
Over 20 trainees, from undergraduate students to postdoctoral fellows, will help develop and train the platform during the five-year initiative. The timeline spans design and pilot testing through to full evaluation and knowledge sharing. The project is expected to produce new benchmarks for judicial AI, policy guidance for courts and regulators, and a verified open-source platform.
Dahan added that the Conflict Analytics Lab, which he directs, is preparing a 2027 CFI Innovation Fund proposal for a roughly $3.9 million project to build Canada's first dedicated sovereign computing and secure data storage system for legal AI. In total, NSERC's 2026 Discovery Research Program granted over $14 million to Queen's University researchers.
Why this matters for legal professionals
Courts are already encountering AI-generated filings, and the tools most litigants use were not built for legal reasoning. This project signals a shift toward AI systems that make their reasoning visible and auditable - a feature that judges, clerks, and opposing counsel can use to assess submissions rather than guess at their origins. The open-source nature of the framework also means the outputs may become available for adaptation across jurisdictions, giving legal professionals a direct line of sight into how AI reaches conclusions in court-facing work.
Your membership also unlocks:
